Energetic island design optimization. A case study using sweep parameter search and variable neighborhood search techniques

This work considers the problem of optimizing the energy management within an Energetic Island by modelling the parameters that configure its components. Such components are typical ones found in renewable energy systems and Smart Grids systems: a power storage system, on demand energy generators and different models for energy generation and load. The optimization consists of minimizing two key values that defines the correct behaviour of the energy management. One is the amount of energy that the customers require (load) but the system is not able to provide. The other one is the energy that is produced by the generators and it could not be consumed. This work considers that an Energetic Island design is optimal if none of these factors reach considerable values while the size and number of is components is controlled; which means a controlled economic costs. Two optimization techniques have been considered. First one performs several parameter search stages through the Parameter Sweep method. Each stage is configured from the best parameter values obtained in the previous one. The other technique is currently at an experimental phase. It consists of the Variable Neighbourhood Search, a well-known method appropriate for finding optimal values within a huge search space.